Over the past nine years we have developed a unique methodology for studying brain structure/function relationships, by studying 708 patients with acute or evolving left hemisphere stroke, initially within 24 hours of onset of symptoms with diffusion-weighted imaging (DWI), perfusion weighted imaging (PWI), and with batteries of language tasks. We have then studied them longitudinally with both imaging and language tests to identify changes in language that correlate with changes in blood flow in the brain. This work has led to novel insights into both the anatomy of the neural networks that support language and how these networks change over time. The networks are sometimes restored to their normal structure in response to restored blood flow and are sometimes reorganized, such that new areas assume the function of the damaged areas during the first year of recovery. In this renewal we aim to combine our unique methodology of studying the neural bases of language (evaluating changes in perfusion with concurrent changes in language performance) with both task-related and resting state functional connectivity MRI (rsfcMRI) to investigate the neural basis of recovery of naming, spelling, lexical-semantics, and sentence comprehension over the first year after stroke. We believe that by combining careful analysis of the changes in cerebral blood flow (in response to stroke or stroke treatment) with the changes in activation (measured by the hemodynamic response), changes in resting state connectivity (measured by correlations in the hemodynamic response in various areas at rest), and changes in anatomical connectivity, we will develop a better understanding of language recovery over the first year after stroke. Task-related fMRI and rsfcMRI provide distinct windows on recovery, and have different advantages. For example, rsfcMRI does not require any response and can be obtained in patients with even severe language deficits. Task-related fMRI reveals brain areas that can be recruited in various activities. Aim 1 will identify changes in neural activation associated with naming, spelling, lexical-semantics, and sentence comprehension at Day 3, Week 2, Month 6, and Year 1 after stroke onset in patients with infarcts and/or hypoperfusion involving left thalamus, other purely subcortical areas, or any one of three language cortex regions of interest (ROIs): inferior frontal gyrus (IFG), superior temporal gyrus (STG), or fusiform gyrus (FG). Aim 2 will identify, in the first three days of stroke, language and neural (imaging) predictors of subsequent language recovery. Aim 3 will identify imaging biomarkers of language recovery at Months 6 and 12. It is hoped that this multidimensional model of recovery will serve in the future as the basis for targeted brain-based interventions for aphasia, which require an understanding of the anatomy of language networks, as well as the extent and timing of how these networks reorganize after injury.